DOI: https://doi.org/10.21203/rs.3.rs-2388971/v1
Background: Maternal and neonatal mortality in Uganda remain persistently high. While utilisation of and utilisation of maternal health services has been shown to reduce the risk of maternal death, little is known about the inequalities in utilisation of maternal health services. This study examined the inequalities in utilisation of maternal health services between 2006 and 2016 to draw implications for achieving universal health coverage.
Methods: We used the Uganda Demographic Health Survey 2006, 2011 and 2016 to analyse inequalities in utilisation of antenatal care (ANC4+), skilled birth attendance (SBA), postnatal care (PNC) and a package of maternal health services. Equity ratios, concentration curves, concentration indices and regression analysis were used in the estimations.
Results: Inequalities in utilization of single and a package of maternal health services reduced between 2005 and 2016, but remained pro-rich. Inequalities in utilisation of package of maternal health services were greater than for a single service. Women from the richest quintile were 4 times more likely to receive a package of care compared to the poorest women, but were just 1.5 times more likely to receive ANC4+ than those in the poorest quintile. In 2006 women in urban areas were 2.6 times more likely to receive a package of all three maternal health services than their rural counterpart and they had a relative advantage of 23.4% to utilize skilled birth delivery than the poorest women. Each additional year of schooling and living in urban areas was associated with 1.2 and 1.6 percentage point increase in utilisation of a package of care respectively. Wealth, education and living in urban areas were positively associated with utilisation of all maternal healthcare.
Conclusion: Declining inequalities in utilisation of maternal healthcare reflect a move towards achieving universal health coverage in Uganda. Pro-rich, education and urban-biased inequalities, imply the need for targeted interventions for the poor, less educated and rural women. Targeted voucher schemes, free distribution of birth kits for poorer and rural women, community-level mobilization to improve uptake of postnatal care, and promoting women’s education and incomes are feasible interventions to improve utilisation of maternal health services and equity.
Uganda remains one of the high-burden countries with high absolute numbers of maternal deaths (WHO, 2015) in sub-Saharan Africa (Orobaton, et al, 2016). Skilled birth attendance has improved to 74% in 2016 from 42% in 2006. However, there remain rural-urban gaps, with 70% of women in rural areas accessing skill-births compared to 88% for urban women. During pregnancy, urban women were twice as likely to see a doctor (17%) than rural women (8%) (UBOS 2018). Similarly, the rate of stillbirths and neonatal mortality equally remain high (Kujala, et al 2017). While Uganda’s maternal mortality ratio (MMR) declined from 687 to 343 per 100,000 live births between 1990 and 2015 (Alkema, et al 2016; Benova el at 2018; MoH, 2019), this is far from the Sustainable Development Goals (SDG) target of reducing the MMR to 70 per 100,000 live births. It also lags behind the Health Sector Development Plan, 2014/15 -2019/2020 national targets of 211 per 100,000 live births. More recent statistics show that despite the prioritization of maternal health services during the HSDP I, there was only a 4%-point increase in the proportion of pregnant women who had four or more antenatal care visits (ANC4+) from 38% in 2015/16 to 42% in 2019/20 which remains short of the HSDP target of 47.5% (MOH, 2020). Over the HSDP-I period, facility-based deliveries improved only slightly from 55% in 2015/16 to 59% in 2019/20 against a target of 89%, while Neonatal Mortality rate (NMR) has stagnated at 27/1000 live births against a target of 10/1000 live births (MOH, 2020).
Antenatal care (ANC) and skilled birth attendance are believed to prevent maternal and perinatal deaths (Campbell, et al 2006; Carroli, et al 2001) and their coverage is routinely used to monitor progress towards improving maternal and neonatal health outcomes. Campbell et al (2006) show that strategies to reduce the burden of maternal mortality in developing countries have proved to be among the most successful efforts to address a specific cluster of causes of death. These strategies included effective intrapartum-care backed up by access to referral-level facilities. Some developing countries such as Indonesia, Guatemala, and Brazil reduced the risk of maternal death by 90–99%. Previous MMR of 1000 per 100 000 livebirths or greater risk of maternal mortality seen in some of these countries, has been reduced to as low as 10 per 100 000. Carroli et el 2001 show that in situations of high maternal mortality, opportunities for pregnant women to access health services such as antenatal care, facility based delivery and management of risk factors at child birth (haemorrhage or obstructed labor) are beneficial.
Previous studies in Uganda have explored factors associated with utilization of maternal health services in Uganda. Coverage with any ANC over the last two decades remained high at over 90%, less than 50% of the women went for four or more ANC visits. While 60% of the women had 4 or more ANC visits, in 2016 only 29% had ANC visits in the first trimester (Benova, et al 2018). Women with higher levels of education were more likely to access early ANC, health facility delivery and early PNC (Atuhaire et al 2020; Ndugga et al 2020; Bbaale 2011). For example, women with post-secondary education had a 33% chance of using professional childbirth care than the non-educated (Bbaale 2011). Similarly, being in higher wealth status as well as residing in urban area was associated with improved utilization of early PNC (Ndugga et al 2020). In Ethiopia, mothers with secondary or higher level of education, residing in urban area and high wealth status and working women had higher odds of delivering at health facilities (Fedaku et al 2019). These previous studies underscore the importance of assessing the utilization of maternal health services across socio-economic groups to understand the nature and trend of inequalities in a single and a package of care.
To achieve Universal Health coverage, it is important that access to and utilisation of maternal health services improves in an equitable manner. There is however limited empirical evidence on utilisation of maternal health services across socioeconomic groups and how this has been changing over time. Understanding the nature and extent of inequities in utilisation of maternal health services over time is important in designing interventions to move towards universal health coverage (UHC), including promoting those targeting particular groups. This study examines the inequities in utilisation of antenatal care (ANC), health-facility assisted births and post-natal care across socioeconomic groups over the period 2005 and 2016 using data from the Uganda Health and Demographic Surveys (UDHS).
The aim of this paper is to examine the extent of inequalities in utilisation of at least four ANC visits (ANC4+), quality ANC, skilled- birth attendance (SBA), post-natal care (PNC) and a package of services across socio-economic groups and between rural and urban women overtime. It also examines the relationship between a mother’s socioeconomic status and utilisation of maternal health services. We used Uganda Demographic and Health Survey (UDHS) data for 2006, 2011 and 2016 collected by the Uganda Bureau of Statistics (UBOS). The UDHS 2006 included a representative sample of 9,864 households selected in two stages. In the first stage, 321 clusters were selected from among a list of clusters sampled in the 2005–2006 Uganda National Household Survey (UBOS, 2006). In the second stage, households in each cluster were selected based on a complete listing of households. In the UDHS 2011, a representative sample of 10,086 households was selected in two stages. In the first stage, 404 EAs were selected from among a list of clusters sampled in the 2009/10 Uganda National Household Survey (2010 UNHS). In the second stage, households in each cluster were selected based on a complete listing of households. The UDHS 2016 used the sampling frame of the Uganda National Population and Housing Census (NPHC), conducted in 2014. A representative sample of 20,880 households (30 per EA or EA segment) was randomly selected.
One of the modules in the UDHS questionnaire collects data on all women aged 15– 49 years of age who had had a birth in the last 5 years. These data include fertility and family planning, maternal and child health as well as details on antenatal and delivery care and other socio-demographic characteristics. Mothers are asked about utilization of maternal health services during the pregnancy for their most recent live birth in the 5 years preceding the survey. The total sample of women aged 15–49 years covered in UDHS 2006, 2011, and 2016 was 8,531, 9,247, and 18,506 respectively (UBOS 2006, 2012 and 2018). All women aged 15–49 who were either permanent residents of the selected households or visitors who stayed in the household the night before the survey were eligible to be interviewed. Details about the sampling procedure for these UDHS studies are reported under respective reports (UBOS 2006, 2012 and 2018).
Inequalities in utilisation of maternal health services are estimated using the theoretical, analytical and methodological approaches for estimating inequities in health and healthcare as proposed by Wagstaff et al, 1991; Culyer and Wagstaff, 1993 and as applied in many previous studies on maternal health (Atuhaire, et al., 2020; Ndugga et al 2020; Dey et al 2020; Fedaku et al 2019; Benova et al 2018, Mohan et al 2015, and Bbaale 2011). The equity analysis in this paper is based on the egalitarian theory of justice as proposed by Rawls (1991), which is the most used in empirical analysis of inequities. Egalitarian theory provides that all social groups are in equal need of a desired healthcare intervention and any distribution which deviates from this principle is considered to be inequitable. Based on the egalitarian theory, each pregnant woman deserves to access the required optimal ANC visits (4 or more) and to give birth from a formal healthcare facility, assisted by a skilled health worker. Similarly she deserves to use postnatal care during the six weeks after birth. (WHO 2013).
We estimate equity ratios, concentration curves, the corresponding concentration indices and regression analysis to examine the nature and extent of socio-economic inequalities in use of a single and a package of maternal health services. Regression analysis is used to explore the factors associated with utilisation of maternal health services. The focus of the regression analysis is to examine how the covariates of socio-economic status indicators of interest to understand the level of inequality. The measures of socio-economic status used in this study are wealth quintile, place of residence-(rural/urban) and education level of the mother.
An equity ratio is estimated as a ratio of the level of utilisation for the top-most quintile to the bottom (poorest) quintile. A ratio greater than 1 represents pro-rich inequalities. We used p-values to examine the level of statistical significance of the estimated indices. The nature and extent of inequality is dependent on the magnitude and sign of the estimated Concentration Index and the p-value at 10%, 5%, and 1% level of confidence.
In addition to focusing on use of any ANC and ANC4+, we also analysed inequalities in quality ANC. Since utilization of ANC is already high (96% for any ANC and 58% for ANC4+) (UBOS 2016) it is important from a public health perspective to examine the quality of ANC women receive. An index for quality ANC was developed based on 8 parameters as recommended by the WHO which comprise the components of ANC visits. These include: a woman weighed, blood pressure measured, urine sample taken, blood sample taken, told about pregnancy complications, given iron supplements, given drugs for intestinal parasites, given SP for malaria prophylaxis
The index was constructed by considering a simple sum of the parameters. A woman who received all components was given a score of 8 and the one who received just one of the 8 was assigned a score of 1. A cut-off point for receiving quality ANC was fixed at a score of 4. Quality ANC is a binary variable taking on 0 for scores less than 4 and 1 otherwise.
Table 1 presents summary statistics of the key socio-demographic characteristics of the sample. Most women had low levels of education with an average years of schooling of 4.4 in 2006 which slightly increased to 6.3 in 2016. Their partners were however better educated with 8 years of schooling on average. Most women and their partners were engaged in agricultural production, which is consistent with the structure of the Ugandan economy where over 70% of the population is involved in agriculture. There were more women in the bottom 2 wealth quintiles than in the top quintiles. Most women lived in rural areas with only 13% in urban areas in 2006 increasing to 23% in 2016. This is consistent with national statistics where 27% of residents are urban-based (UBOS 2022).
| Characteristic | 2006 | 2011 | 2016 |
|---|---|---|---|
| Woman age at last birth (mean) | 28.7 | 28.76 | 26.61 |
| Woman years of schooling | 4.413 | 5.365 | 6.303 |
| Partner years of schooling | 8.003 | 7.611 | 8.109 |
| Household size | 6.571 | 6.362 | 6.002 |
| Woman occupation: Professional/technical/managerial | 2.8% | 4.3% | 8.2% |
| Woman occupation: Sales and services | 11.3% | 19.9% | 14.3% |
| Woman occupation: Agricultural/fishery | 72.0% | 55.5% | 43.6% |
| Woman occupation: (Un)Skilled manual | 5.0% | 15.2% | |
| Woman occupation: No work or household chores | 8.9% | 20.3% | 18.7% |
| Partner occupation: Professional/technical/managerial | 7.6% | 7.2% | 12.3% |
| Partner occupation: Sales and services | 13.2% | 23.3% | 10.2% |
| Partner occupation: Agricultural/fishery | 55.7% | 65.2% | 47.0% |
| Partner occupation: (Un)Skilled manual | 18.2% | 27.5% | |
| Partner occupation: No work or household chores | 5.3% | 4.3% | 2.9% |
| Wealth index: Lowest | 21.3% | 21.3% | 20.9% |
| Wealth index: Second | 21.7% | 20.7% | 20.4% |
| Wealth index: Middle | 19.6% | 19.4% | 18.9% |
| Wealth index: Fourth | 19.0% | 18.0% | 18.3% |
| Wealth index: Highest | 18.3% | 20.7% | 21.5% |
| Household possesses radio/TV: Yes | 61.0% | 66.6% | 59.6% |
| Residence: Urban | 13.2% | 16.1% | 23.1% |
| Number of Observations | 8531 | 9247 | 18506 |
The level of utilisation to maternal health services varied depending on the type of service. While about 98% of the women accessed at least one ANC visit, just 60% had received four or more visits in 2016 as shown in Table 2. Majority of women were starting ANC visits fairly late with only 29% starting within their first gestation period in 2016 compared to 16.5% in 2006. Utilisation of a single MHS (i.e. ANC4+, skilled birth attendance, and PNC) was higher than for a continuum of care. (‘i.e.’ 2 or all 3 maternal health services).
| Characteristic | 2006 | 2011 | 2016 |
|---|---|---|---|
| ANC visits: 4 + visits | 47.2% | 47.5% | 59.9% |
| ANC visits: 1–3 visits | 48.1% | 48.0% | 38.1% |
| ANC visits: 0 visits | 4.6% | 4.5% | 2.0% |
| Gestation age at first visit: ≤ 3 months | 16.5% | 20.7% | 29.1% |
| Gestation age at first visit: 4–6 months | 62.5% | 62.3% | 61.6% |
| Gestation age at first visit: 7 + months | 16.4% | 12.5% | 7.3% |
| Skilled delivery (SBA): Yes | 45.2% | 60.4% | 76.2% |
| Postnatal care: Yes | 20.2% | 37.7% | 54.0% |
| 4 ANC visits + SBA: Yes | 26.7% | 32.8% | 49.0% |
| 4 ANC visits + SBA + PNC: Yes | 12.2% | 18.9% | 33.6% |
| Number of Observations | 8531 | 9247 | 18506 |
| Source: Estimations from UDHS, 2006, 2011 & 2016 datasets | |||
The inequalities in utilisation of maternal health services are estimated based on the equity ratio, concentration index (and corresponding concentration curve) and a multivariate regression-based analysis. The estimated equity ratios for utilisation of ANC4+, SBA and a continuum of maternal health services are shown in Table 3. These ratios are derived from the estimated percentage utilisation of maternal health services by the top-most quintile versus the bottom quintile. The equity ratio shows the number of times the richest group is better off in terms of accessing care compared to the poorest group.
| Equity ratio (Richest/Poorest) | 2006 | 2011 | 2016 |
|---|---|---|---|
| 4 ANC visits: Equity ratio | 1.451 | 1.376 | 1.234 |
| SBA: Equity ratio | 2.625 | 1.987 | 1.437 |
| PNC: Equity ratio | 2.762 | 2.030 | 1.474 |
| 4 ANC visits + SBA: Equity ratio | 3.344 | 2.571 | 1.602 |
| 4 ANC visits + SBA + PNC: Equity ratio | 4.333 | 3.176 | 1.824 |
The higher the equity ratio the higher the inequality is against the poor (also referred to as pro-rich inequality). From Table 3 the nature of inequality in utilisation of maternal health services (both single and a combination of services) is pro-rich. For example in 2016, women in the top most quintile were 1.8 times more likely to use a package of all three maternal health services than their counterparts in the poorest quintile. This implies that the poor women are disadvantaged regarding utilisation of maternal health services. However, the level of inequality has been reducing since 2006.
Inequalities were also assessed based on the place of residence to show the rural/urban differences in accessing maternal health services. The estimates in Table 4 shows that utilisation of maternal health services was relatively better for women in the urban compared to rural areas for all types of services. The inequality ratios estimated for each type of services shows by how much women in urban areas are more likely to access care than their rural counterparts. For example, in 2006, women in urban areas were 2.6 times more likely to receive a combination of all three maternal health services (ANC4 + SBA + PNC) than their rural counterparts. This reflects inequalities against the rural women. However, there has been a declining trend in the extent of inequalities (differences) in utilisation of maternal health services since 2006. This is a positive result in view of the national and global goal of moving towards universal health coverage in maternal health and across all healthcare services more generally.
| Estimate | 2006 | 2011 | 2016 |
|---|---|---|---|
| 4 ANC visits: Rural | 45.3% | 45.7% | 58.3% |
| 4 ANC visits: Urban | 59.7% | 57.0% | 65.2% |
| Inequality (Urban/Rural) | 1.318 | 1.247 | 1.118 |
| SBA: Rural | 39.7% | 55.0% | 71.9% |
| SBA: Urban | 83.3% | 89.7% | 90.6% |
| Inequality (Urban/Rural) | 2.098 | 1.631 | 1.26 |
| PNC: Rural | 17.5% | 33.4% | 49.9% |
| PNC: Urban | 39.9% | 60.2% | 67.4% |
| Inequality (Urban/Rural) | 2.28 | 1.802 | 1.351 |
| 4 ANC visits + SBA: Rural | 22.8% | 29.0% | 45.6% |
| 4 ANC visits + SBA: Urban | 52.4% | 53.0% | 60.0% |
| Inequality (Urban/Rural) | 2.298 | 1.828 | 1.316 |
| 4 ANC visits + SBA + PNC: Rural | 10.2% | 15.4% | 30.4% |
| 4 ANC visits + SBA + PNC: Urban | 26.1% | 36.9% | 44.5% |
| Inequality (Urban/Rural) | 2.559 | 2.396 | 1.464 |
| Source: Computations based on UDHS data, 2006, 2011 and 2016 | |||
We estimated the concentration index (and corresponding concentration curves) to counteract the limitations of the equity ratio which does not consider the distribution across all SEG. The concentration curves (based on UDHS 2016 data) for ANC4 + and SBA are shown in Fig. 1 below, while those for a combination of maternal health services are given in Fig. 2. Since the health variable of interest is a health good (utilisation of maternal health services), the concentration curve lying below the line of perfect equality (the diagonal) implies inequality against the poor. The extent of divergence from the line of perfect equality shows the extent of the existing inequalities.
[Figure 1 here]
[Figure 2 here]
The inequalities in utilisation of skilled-birth delivery were larger than for ANC4+. Likewise, the inequalities in utilisation of a combination of ANC4 + and SBA were larger than for a single maternal health service (Fig. 2). The nature of inequalities in maternal health services based on the concentration curves are consistent with the estimated equity ratios.
The inequality estimate based on concentration indices are shown in Table 5 below. The positive sign shows that the inequalities are pro-rich (or against the poor).
| 2006 | 2011 | 2016 | |
|---|---|---|---|
| ANC visits: 4 + visits | 0.068(0.010)*** | 0.068(0.011)*** | 0.042(0.007)*** |
| Skilled delivery: Yes | 0.194(0.012)*** | 0.132(0.009)*** | 0.077(0.005)*** |
| Postnatal care: Yes | 0.217(0.021)*** | 0.147(0.015)*** | 0.085(0.008)*** |
| 4 ANC visits + SBA: Yes | 0.243(0.017)*** | 0.185(0.014)*** | 0.096(0.008)*** |
| 4 ANC visits + SBA + PNC: Yes | 0.304(0.027)*** | 0.251(0.023)*** | 0.130(0.011)*** |
Standard errors in parentheses * p < 0.05, ** p < 0.01, *** p < 0.001
Inequalities in a single Maternal health service and a continuum of care have been reducing over the period 2006 to 2016 as shown by the decline in the magnitude of the concentration indices over this same period but remain pro-rich. Relative to utilisation of a single package of maternal health services (e.g. ANC4 + or SBA), inequalities are greater in accessing a combination of services (“i.e.” ANC4 + SBA + PNC, & ANC4 + SBA). Inequalities are greater for jointly accessing all three maternal health services (ANC4 + SBA + PNC) and just two of the full packages of comprehensive maternal healthcare (i.e. ANC4 + SBA). The results of the concentration indices, which consider the distribution across all population groups are consistent with the estimated equity ratios, which also depicted a reduction in the level of inequality overtime.
This section presents findings on the predictors of utilisation of quality ANC skilled birth attendance, PNC and a combination of maternal health care. The SES variables of interest in this analysis were place of residence, wealth status and education of the mother. The socio economic inequalities in use of quality ANC are shown in Table 6. The magnitude and nature of inequality are reflected by the size and the sign of the estimated coefficient respectively. The estimates are margins which show the percentage changes in the dependent variable to a percentage change in the SES indicator. The results reported in Table 6 to 10 are for the SES variables of interest (‘i.e.’ wealth, education of a woman and residence). Detailed results including cofounders are shown in the supplementary file.
Education was positively associated with utilisation of quality ANC over the time period. An additional year of schooling was associated with a 2.5%age point’s increase in utilisation of quality ANC in 2006 and 1.6%age point increase by 2016. Wealthier women were more likely to receive quality ANC compared to those in the lower wealth quintiles. In 2006, women in the lowest wealth quintile were 21.3%age points less likely to receive utilisation of quality ANC compared to those in the top most quintile slightly reducing to 20.0%age points by 2016. In 2006, women residing in urban areas were 38.7%age points more likely to access quality ANC services compared to the rural areas. However this relative advantage reduced to just 8.1%age points by 2016 which shows significant decline in inequalities.
| 2006 | 2011 | 2016 | |
|---|---|---|---|
| Woman years of schooling | 0.025 (0.006)*** | 0.029 (0.005)*** | 0.016 (0.003)*** |
| Wealth index: Lowest | -0.213 (0.071)*** | -0.026 (0.068) | -0.200 (0.052)*** |
| Wealth index: Second | -0.220 (0.064)*** | -0.163 (0.057)*** | -0.246 (0.050)*** |
| Wealth index: Middle | -0.231 (0.066)*** | -0.206 (0.055)*** | -0.219 (0.043)*** |
| Wealth index: Fourth | -0.177 (0.059)*** | -0.144 (0.050)*** | -0.143 (0.038)*** |
| Residence: Urban | 0.387 (0.075)*** | 0.277 (0.074)*** | 0.081 (0.038)** |
| R squared | 0.221 | 0.303 | 0.296 |
| Observations | 4947 | 4868 | 10263 |
Standard errors in parentheses * p < 0.10, ** p < 0.05, *** p < 0.01
Results for the estimated inequalities in utilisation of skilled birth delivery (Table 7) show that a unit increase in the years of school for the mother increased the likelihood of accessing skilled-birth delivery by 1.9%age, 1.5%age and 1.1%age points. Likewise, in 2006 women in the lowest wealth quintile were 15.1% age points less likely to be delivered by a skill birth attendant compared to women in the richest SEG. The relative advantage of the women in the richest SEG to access skilled birth delivery increased to 23.4% age points in 2011 before reducing to 10.8%age points, but the distribution remained inequitable against the poor. The relative advantage between women in the richest group and those in the 4th, 3rd and 2nd was correspondingly lower, and declined between 2006 and 2016. That is to say, the advantage of the women in the richest group to access skilled-birth delivery relative to those in the 4th SEGs is lower than that relative to the women in the 3rd quintile, and so on. In 2006, women in urban areas were 13.5 percentage points more likely to have a birth assisted by a skilled health worker compared to those in rural areas. This relative advantage reduced to 9.4 percentage points in 2011 and further to 4.8 percentage points in 2016, depicting a declining trend in inequalities. Estimates for the full regression model are in the supplementary file.
| 2006 | 2011 | 2016 | |
|---|---|---|---|
| Woman years of schooling | 0.019 (0.003)*** | 0.015 (0.003)*** | 0.011 (0.002)*** |
| Wealth index: Lowest | -0.151 (0.042)*** | -0.234 (0.038)*** | -0.108 (0.024)*** |
| Wealth index: Second | -0.142 (0.037)*** | -0.180 (0.031)*** | -0.124 (0.019)*** |
| Wealth index: Middle | -0.138 (0.034)*** | -0.166 (0.029)*** | -0.078 (0.017)*** |
| Wealth index: Fourth | -0.072 (0.032)* | -0.136 (0.028)*** | -0.040 (0.014)** |
| Residence: Urban | 0.135 (0.044)** | 0.094 (0.025)*** | 0.048 (0.014)*** |
| R squared | 0.206 | 0.157 | 0.120 |
| Observations | 4947 | 4868 | 10263 |
Standard errors in parentheses * p < 0.05, ** p < 0.01, *** p < 0.001
Unlike inequality estimates for quality ANC (Table 6) and SBA (Table 7), inequalities in utilisation of postnatal care had an increasing trend against the poor between 2006 and 2016. The findings in Table 8 show that more educated mothers were more likely to access PNC compared to their less educated counterparts. Whereas an additional year of schooling increased the likelihood of receiving PNC by 1.1 percentage points, this increased to 1.7 percentage points in 2011 and slightly reduced to 1.4 percentage points in 2016 reflecting an increase in inequalities based on the education level of the woman. Compared to women in the top-most wealth quintile, women in the lower wealth quintiles were less likely to access postnatal care with an increasing trend in inequality over the 10 year period. Likewise women in urban areas were more likely to receive PNC than those in the rural areas.
| 2006 | 2011 | 2016 | |
|---|---|---|---|
| Woman years of schooling | 0.011 (0.003)*** | 0.017 (0.003)*** | 0.014 (0.002)*** |
| Wealth index: Lowest | -0.049 (0.033) | -0.122 (0.044)** | -0.081 (0.030)** |
| Wealth index: Second | -0.078 (0.031)* | -0.131 (0.040)** | -0.133 (0.028)*** |
| Wealth index: Middle | -0.087 (0.032)** | -0.157 (0.039)*** | -0.115 (0.026)*** |
| Wealth index: Fourth | -0.073 (0.028)** | -0.119 (0.039)** | -0.088 (0.024)*** |
| Residence: Urban | 0.043 (0.030) | 0.063 (0.035) | 0.043 (0.022) |
| R squared | 0.121 | 0.107 | 0.075 |
| Observations | 4947 | 4868 | 10263 |
Standard errors in parentheses * p < 0.05, ** p < 0.01, *** p < 0.001
The findings on the level and nature of inequalities in accessing a combination of maternal health services are reported in Tables 9 (ANC4 + and SBA) and Table 10 (ANC4+, SBA and PNC).
From Table 9, being more educated, in a higher wealth quintile as well as living in the urban area were positively associated with accessing a combination of ANC4 + and skilled birth attendance. This reflects inequalities against the women who were socio economically disadvantaged. However the inequalities declined between 2006 and 2016, reflecting a positive trend towards UHC.
| 2006 | 2011 | 2016 | |
|---|---|---|---|
| Woman years of schooling | 0.013 (0.003)*** | 0.007 (0.003)** | 0.010 (0.002)*** |
| Wealth index: Lowest | -0.164 (0.036)*** | -0.200 (0.041)*** | -0.094 (0.026)*** |
| Wealth index: Second | -0.152 (0.034)*** | -0.154 (0.034)*** | -0.088 (0.024)*** |
| Wealth index: Middle | -0.161 (0.031)*** | -0.139 (0.033)*** | -0.076 (0.022)*** |
| Wealth index: Fourth | -0.137 (0.031)*** | -0.108 (0.031)*** | -0.039 (0.022) |
| Residence: Urban | 0.054 (0.034) | 0.035 (0.029) | 0.027 (0.018) |
| R squared | 0.187 | 0.156 | 0.153 |
| Observations | 4947 | 4868 | 10263 |
Standard errors in parentheses * p < 0.05, ** p < 0.01, *** p < 0.001
Unlike the inequalities in accessing ANC and SBA which declined over-time, those for a combination of all three services increased between 2006 and 2016 as shown in Table 10. Women who were more educated, of a higher wealth status as well as living in the urban area were more likely to access a combination of all three maternal health services compared to their counterparts.
| 2006 | 2011 | 2016 | |
|---|---|---|---|
| Woman years of schooling | 0.008 (0.002)*** | 0.010 (0.002)*** | 0.012 (0.002)*** |
| Wealth index: Lowest | -0.063 (0.026)** | -0.135 (0.034)*** | -0.096 (0.026)*** |
| Wealth index: Second | -0.067 (0.025)*** | -0.135 (0.032)*** | -0.116 (0.025)*** |
| Wealth index: Middle | -0.081 (0.024)*** | -0.132 (0.032)*** | -0.120 (0.022)*** |
| Wealth index: Fourth | -0.067 (0.024)*** | -0.105 (0.029)*** | -0.079 (0.024)*** |
| Residence: Urban | 0.002 (0.025) | 0.047 (0.028)* | 0.016 (0.019) |
| R squared | 0.134 | 0.128 | 0.104 |
| Observations | 4947 | 4868 | 10263 |
Standard errors in parentheses * p < 0.10, ** p < 0.05, *** p < 0.01
This paper set out to examine the extent of inequalities in utilisation of at least four ANC visits (ANC4+), including quality ANC, skilled-birth attendance (SBA), post-natal care (PNC), and a continuum of services across socio-economic groups and between rural and urban women overtime. We used the common analytical approaches for estimating inequities in health and healthcare as proposed by Wagstaff et al, 1991; Culyer and Wagstaff, 1993 to estimate equity rations, concentration indices and regression analysis based on UDHS data for 2006, 2011 and 2016. The methodological contribution of this paper were three fold: first to analyse the trend in inequalities over a 10 year period while the previous studies were looking at single data points. Secondly we estimate utilisation and inequalities in a package of care by considering those women who had two (ANC and SBA) or all three (ANC + SBA + PNC) maternal health services. Thirdly we estimate inequalities in utilisation of quality ANC based on a weighted index of quality parameter as explained in the methods section. Empirical evidence on utilisation of quality ANC is limited, although important in situations when overall utilisation of any ANC care is very high as the case in Uganda. Assessment of levels of utilisation, inequalities and quality of care are important considerations for moving towards UHC.
Majority of the women were starting ANC visits fairly late, with only 29% of the women in 2016 who initiated ANC visits within their first trimester (within first 3 months of pregnancy) as recommended by health professionals. Starting the ANC visits late during pregnancy exposes the mother and baby to potential risks which could be addressed if detected early. It also reduces the quality of ANC services the pregnant mother can have since some services within the full package of services may not be provided after the 3rd trimester (e.g. SP for malaria in pregnancy), and she therefore misses the opportunity to receive vital services such as vitamin A and C supplementation and HIV/AIDS testing (to prevent mother to child transmission). In order to have a sufficient number of visits to ANC clinics (4 or more visits by 2016, revised to 8 or more in 2018 by the WHO) a woman should start their ANC visits in the first 3 months of pregnancy.
Utilisation of maternal health services has improved since 2006, although still remains below the HSDP (2021–2025) targets, particularly SBA and PNC. The HSDP target for SBA is 89% against the current level of 76%, implying that interventions to enable women deliver assisted by a skilled health worker need to be intensified. These include ensuring that the maternity wards in health facilities are functional with required supplies and health staff, social mobilisation by the village health teams at community level to encourage and refer pregnant women to health facilities; maternal health education campaigns, and addressing the financial and other health-care seeking barriers at the household and community level. Utilisation of a package of MHS is far below use of a single maternal health service. This implies there are women who may get the required number of ANC visits but do not deliver with the assistance of a skilled health worker – exposing them to post-partum complications, which could be avoided, or properly handled by a skilled birth attendant. Likewise, some women who receive the required number of ANC visits and also deliver from a health facility, do not receive the vital checks post-delivery to assess the health of the mother and new baby (PNC). Failure to receive such important health services could lead to severe illnesses and complications which are preventable if the mother received a package of care during pregnancy.
Overall, however, utilisation of maternal health services improved between 2006 and 2016, including among women with lower socioeconomic status. This could be attributed to a range of interventions implemented by government and partners over this period. These interventions, included those by MOH and partners, such as the stop malaria project which was distributing LLINs during ANC visits, leading to drastic increase in use of ANC within the SMP implementing districts. The PROFAM Social franchising programme in Uganda to improve access for maternal services within the private sector through health-worker trainings, supply of medical equipment and quality improvement programme, also increased utilisation of MHS at private health facilities as evidenced by the evaluation by Haemmerli, et al (2018). The Mama Kit intervention in the Ministry of Health also increased utilisation of SBA for poorer women by reducing providing delivery kit to women during their ANC visits to public health facilities. Other project-based interventions include those by “Saving mothers Living Life” Marie Stopes International, among others.
Whereas the inequalities in utilisation of maternal health services have declined since 2006, they remain pro-rich. The decline in inequalities is a positive trend in relation to achieving equitable utilisation of maternal health services, which is a key factor in moving towards universal health coverage in the country. However, wealthier, more educated and urban-based women have an advantage in using maternal health services compared to their poor, less educated and rural-based counterparts, who are also the majority in Uganda. These findings are similar to those by Atuhaire, et al 2020, Ndugga et al 2020, Benova et al 2018, and Bbaale 2011 studies on Uganda who found that a woman’s wealth status, years of schooling education and living in urban areas were positively associated with likelihood of using maternal health services. Empirical work by Weitzman (2017) in Peru, Sepehri, et al (2008) in Vietinam; Yadav, et al (2021) and Yadav& Jena (2022) for India and a study on utilisation of maternal health services in five African countries by (Dimbuene, et al 2018) reported similar findings. Our study adds value to the exiting knowledge by assessing the trends in inequalities, as well as considering utilisation of a package of MHS, which previous studies did not explore. Thus, making maternal health services accessible to poorer, less educated and rural-based women remains a priority in order to achieve UHC.
Utilisation of a package of all three maternal health services (ANC4+, SBA and PNC) remain quite low and significantly concentrated among the richer, more educated and urban based women. This is particularly so because of the very few women that access postnatal care after delivery. This could be because the cost of care is more affordable to such women, they have better access to health facilities in the urban areas, and have better access to information and knowledge about the benefits of MHS. A dearth of studies have shown that better educated women have greater capacity to access and process available information on healthcare and take action to access care (Shahabuddin et al, 2015; Wado, 2017; Dimbuene, et al, 2018;Yadav, et al 2021; Yadav & Jena, 2022). They also tend to become more aware of the potential dangers of not accessing necessary care. Similarly, wealthier women have the capacity to meet the financial cost of accessing care, such as transport cost, which remain a barrier to healthcare access. The finding that more educated women with a higher socioeconomic status were more likely to receive MHS, both within the rural and urban areas (Sepehri, et al, 2008; Kalule-Sabiti, et al 2014; Shahabuddin et al 2015; Weitzman, 2017; Dimbuene, et al ,2018; Barman, et al, 2020; Yadav, et al 2021;). This suggests that over the long-term, increasing education opportunities for girls and young women, and interventions to improve household incomes, has a direct effect on improving utilisation of maternal health services and other forms of healthcare generally. Increasing utilisation of care, and in an equitable manner are important for the attainment of UHC.
Utilisation of any ANC and ANC4 + services has improved since 2006. Despite this improvement women accessing quality ANC remain few and or are concentrated among the high income groups. This has implications on maternal health outcomes in the underserved socio-economic groups. Women who were more likely to access quality ANC were those from wealthier quintiles, more educated, and from urban areas. Thus interventions targeted at the poorer and rural based women are necessary to improve overall maternal health outcomes. The literature shows that receiving ANC services early on during the pregnancy, and making sufficient number of ANC visits (ANC4 + revised to ANC8 + by WHO since 2016) are important for better health outcomes during delivery and the postnatal period (Bergsjø, 2001; Kuhnt & Vollmer, 2017). However, the findings in this study showed that for some SEGs the proportion of women who delivered at a health facility (SBA) was greater than those who had received ANC4+. This suggests that some women although delivered at a health facility in the hands of a skilled birth attendant, they had not received the recommended number of ANC visits.
Similar to ANC, there has been a steady decline in the level of inequalities in utilisation of skill-birth deliveries overtime. Moreover, the proportion of the less educated and poorer women accessing skilled-birth deliveries has been increasing overtime, while the level of inequality has been declining. These findings reflect the positive effect of the interventions by government and other partners to improve utilisation of maternal health services across all population groups and have important implications for achieving UHC. The interventions include: government policy to have a functional HC3 with maternity services at sub-county level to increase utilisation of health facilities, improved staffing levels, and reducing stock-outs of essential medicines and health supplies at health facilities, among others. The urban/rural inequalities can be attributed to the imbalances in availability of sources of care between rural and urban areas. There are more healthcare providers for maternal health services (and other healthcare services) in the urban compared to rural areas. From the health system perspective, it will be important to increase sources of care for maternal health services in rural areas, through expanding the public healthcare infrastructure which provides maternal health services, as well as promoting Public-private partnerships initiatives for health care to expand services in areas where the public health system is not well-established to provide needed care. Social mobilisation of women through health campaigns at the community level within the VHT system can help to ensure women start their ANC visits early during pregnancy.
Women who had no ANC visits were less likely to utilise skilled-birth delivery compared to those who had at least one ANC visit. This implies that a woman’s utilisation of ANC during pregnancy highly influences their utilisation of skilled birth delivery. This underscores the importance of providing an integrated package of maternal health services for all pregnant mothers.
There are two limitations of this paper which merit discussion. First, in computing the index for quality ANC, equal weights were given to each indicator, It is plausible to expect that some parameters among the quality indicators assessed (‘i.e.’ a woman weighed, blood pressure measured, urine sample taken, blood sample taken, told about pregnancy complications, given iron supplements, given drugs for intestinal parasites, given SP for malaria prophylaxis) could be much more important than others from a medical perspective.
Secondly, we did not have firm empirical evidence for the cut-off of 50% used to categorise quality ANC. It is possible that there are critical ANC services which must be provided to qualify ANC services provided as being of good quality. We did not have published evidence to inform the basis for the cut-off point used.
Lastly, we did not consider the current WHO recommendation of ANC + 8 in our analysis, but rather ANC4 + since the new policy had not been passed by the time of the UDHS2016. In view of the current WHO recommendation for ANC + 8, it is quite likely inequalities for ANC8 + are much higher than what is estimated in this paper for ANC4+.
Inequalities in utilisation of single and a package of maternal health services have declined since 2006, which is a positive trend for achieving UHC. Overall, utilisation of maternal health services improved between 2006 and 2016, including among women with lower socioeconomic status. The proportionate increase in the women in lower wealth quintiles receiving MH care could be attributed to the range of interventions implemented over the period which focused on improving access for all women, including the poor as discussed above. However, the pro-rich nature of existing inequalities suggests that there is need for targeted interventions to improve access among the poorer, less educated and rural women. Targeted interventions to improve utilisation of MHS for the poor could include voucher schemes to enable poorer women access care, especially for SBA and PNC where inequalities are larger than for ANC4+. Improving utilisation of these services will eventually address the existing inequalities in receiving a package of MH care.
The decreasing inequalities in maternal health services, though remain pro rich, imply that the existing interventions are effective and need to be sustained to further reduce the existing inequalities to equalize access between the poor and the richer women and those in rural and urban areas. Examining the nature and trend in inequalities is important to provide insights into the movements towards achieving UHC in maternal health care delivery and suggest how to address existing inequalities.
Ethics approval and consent to participate.
Not applicable.
The study used secondary data from Uganda demographic Health surveys of 2006, 2011 and 2016.
Consent for publication.
Not Applicable
Availability of data and materials
The datasets analyzed during the current study are available in the U.S. Agency for International Development repository, [PERSISTENT WEB LINK TO DATASETS].
https://dhsprogram.com/data/dataset/Uganda_Standard-DHS_2011.cfm?flag=0
All methods were carried out in accordance with relevant guidelines and regulations.
Competing interests
The authors declare that they have no competing interests.
Funding
The authors declare that there was no funding received to carry out this study.
Authors' contributions
PA conceived the study, analyzed the data and drafted the first version of the manuscript. EEK and JM reviewed the draft and provided substantial input. All authors approved the final version of the manuscript.
Acknowledgements
Not applicable
Authors' information
Graduate student, School of Economics, College of Business and Management Science, Makerere University Kampala Uganda
Phiona Atuhaire
Senior Lecturer, School of Public Health, College of Health Sciences, Makerere University, Kampala Uganda.
Elizabeth Kiracho-Ekirapa
Senior Lecturer, School of Economics, College of Business and Management Science, Makerere University Kampala Uganda.
John Mutenyo